Real-time prediction of lean blowout using chemical reactor network

Abstract

Lean blowout (LBO) of combustion systems is a concern that can cause costly and time-intensive reignition of land-based gas turbines and can affect the rate of descent for aircraft and the maneuverability of military jets. This work explores the feasibility of model-based combustor monitoring and real-time prediction of combustion system proximity to LBO. The approach makes use of (1) real-time temperature measurements, coupled with (2) the use of a real-time chemical reactor network (CRN) model to interpret the data as it is collected. The approach is tested using a laboratory jet-stirred reactor (JSR), operating premixed on methane at near atmospheric pressure. The CRN represents the combustion reactor as three perfectly stirred reactors (PSRs) in series with a recirculation pathway; the model inputs include real-time temperature measurements and mass flow rates of fuel and air. The goal of the CRN is to provide a computationally fast means of interpreting measurements in real time regarding proximity to LBO. The CRN-predicted free radical concentrations and their trends and ratios are studied in each combustion zone. The results indicate that the hydroxyl radical maximum concentration moves downstream as the combustion reactor approaches LBO. The ratio of hydroxyl radical concentrations in the flame zone versus the recirculation zone is proposed as a criterion for the LBO proximity. The model-based process monitoring approach sheds insight into combustion processes in aerodynamically stabilized combustors as they approach LBO.